Vascular damage is an important mechanism of PDT action with many photosensitizers and photosensitizing conditions. The extent of this damage can be expected to vary as a function of photosensitizer type, photosensitizer dosing, drug-light interval, fluence, and fluence rate 
. However, the effect of animal strain on PDT-induced vasoresponse has barely been considered. This led us to investigate how tumor hemodynamics during PDT differed between RIF tumors as a function of their propagation in C3H versus nude mice. Tumors of C3H mice developed more severe ischemia during PDT, and this finding could not be attributed to differences in either photosensitizer uptake or pre-existing oxygen tensions between tumors of the two strains. In contrast, tumor vessels of C3H mice were smaller than those in nudes and demonstrated more tightly controlled hemodynamics, evidenced by the presence and regularity of cyclic blood flow patterns; cyclic patterns were absent or not as well defined in the larger tumor blood vessels of the nude mice. Our data show that these strain-dependent differences in tumor hemodynamics have implications in studies of PDT. These differences may also play a role in response to systemic but less severe vascular insults, such as the vasoconstrictor L-NNA, which led to significant decreases in tumor blood flow in C3H but not nude mice.
A role for mouse strain in the response of tumor blood flow to vasoactive drugs has previously been unclear. A handful of papers have considered this possibility with the conclusion that animal strain did not play a role because differentials in blood flow response between strains were small compared to differences between tumor models 
. Nevertheless, the data showed drug-induced changes in blood flow to be slightly bigger for the same tumor model grown in C3H mice versus nude or SCID animals. Laser Doppler was used to measure tumor blood flow in these studies resulting in a sampled area that was small (~1 mm3
) compared to the tumor-wide average provided by DCS in the present study. Given the known intra-tumor heterogeneity in tumor vascularization and its reactivity these differences in sampling size could readily explain the more clear-cut results of the present study.
The immune system is a factor that should be considered when comparing vascular damage between nude and C3H animals. Propagation of the same tumor model in two different murine strains necessitated that one of theses strains be immunodeficient. Previous reports have shown that the efficiency of PDT is compromised in immunodeficient hosts 
. The fact that C3H mice have an intact immune system, while nude mice are athymic and immune deficient, may account for some of the strain-dependent differences due to PDT. However, it is unlikely that the differences we found were a result of differences in T-cell response to PDT because our studies were limited to changes in blood flow during the course of PDT (30 minute treatment), which is likely earlier than the consequences of T-cell deficiency on PDT outcomes would be realized in the nude animals.
On the other hand, immunological-dependent differences in strain responses could be mediated through its effect on tumor development. Solesvik et al. 
have documented differences in the vascular composition of multiple tumors lines when propagated in nude versus C3H hosts. Specifically, in the tumors of C3H mice, small diameter (~10 µm) blood vessels were significantly longer, in fact sometimes twice as long, as blood vessels of similar diameter in the tumors of nudes. Given the tortuous nature of tumor vascular networks, a vessel can transverse a 2D plane multiple times and the longer that a vessel is the more likely it is that this will happen. Thus, the increased length of small diameter blood vessels in the tumors of C3H animals could contribute to our finding that C3H mice bore tumors with smaller-sized vessels. Solesvik et al. did not consistently detect decreases in vessel diameter in C3H versus nude mice across all tumor models studied, but some did indeed show trends toward decreased diameter in the C3H animals (and the RIF model was not among those investigated).
Potential factors contributing to differentials in vessel size between nudes and C3H may be found in studies of the effects of immunogenicity on tumor-associated fibrosis and interstitial fluid pressure. Nude animals lack the lymphocyte-laden fibrosis associated with tumors in immune competent mice 
. Given the stronger fibrotic reaction in immune competent animals, which could increase cell density surrounding the blood vessels, it comes as no surprise that others report tumor interstitial fluid pressure to be higher in immune competent mice than in the same tumor model when grown in nude animals 
. Our findings of smaller-sized vessels in the C3H versus nude animals would therefore be in agreement with expected differentials in tumor interstitial fluid pressure as a function of immunogenicity. Namely, vessel compression could result from high interstitial fluid pressure in immune competent animals, while lower interstitial fluid pressures in the tumors of immune deficient animals could be permissive of larger-sized vessels. Taken together these data point to vascular structure as a variable to be considered in studies of tumor response to vasomodulation within different murine strains.
Differences in vessel size between the tumors of C3H and nudes could contribute to the strain-dependent differences in vasoresponse due to the effect of vessel size on blood flow. Resistance to blood flow is inversely related to the fourth power of vessel diameter 
, predicting that vasoconstriction of a small vessel would increase flow resistance to a greater magnitude than the same vasoconstrictive insult in larger vessels (Poiseuille's law assuming vessel radius decreases by the same amount in each case). Under the assumptions that the concentration of blood vessels is the same in the two models (data show that the vascular area is the same) and that blood pressure is the same (or changes in the same way) between the models, one can then calculate that vasoconstriction of the smaller tumor blood vessels in C3H mice would lead to larger relative changes in blood flow than would the same amount of vasoconstriction in the larger vessels of nude mice. Indeed, PDT-created decreases in tumor rBF were significantly larger in C3H animals. Moreover, the vasoconstrictor L-NNA significantly decreased tumor blood flow in C3H animals, while resulting in smaller, more variable and overall insignificant decreases in tumor rBF in the nudes.
Our findings of oscillations in tumor blood flow under baseline conditions are consistent with many previous reports of fluctuating blood flow or oxygen levels in tumors 
. Such fluctuations have even been noted in studies of PDT along with the observation, as we made, that they are destroyed by the treatment 
. Factors such as tumor size or vessel maturity (expression of smooth muscle actin, SMA) 
can affect the nature of the fluctuations, but these characteristics did not differ between the tumors of C3H versus nudes in our study: mean (SD) tumor size was 270 (88) mm3
vs. 292 (61) mm3
and SMA expression was 0.77 (0.36)% vs. 0.56 (0.58)% in C3H vs. nudes, respectively. The underlying cause of cyclic blood flow or hypoxia has been attributed to several factors, including vasomotion linked to upstream circulation, the local hemodynamics of blood flow through tortuous tumor vasculature and vascular intussusception from rapid vessel remodeling 
. Our studies do not rule out contributions from the latter two factors, but the fact that flow cycled over the tumor as a whole (i.e. the area over which DCS measured) suggests a less-localized cause for the fluctuations. Vasomotion could provide one explanation. Using window chamber models, others have documented vasomotion to occur when the diameter of tumor-feeding arterioles produced coordinated changes in vessel diameter and blood flow through downstream daughter vessels 
. Looking more generally, we have shown in the present study that the low frequency fluctuations in tumor blood flow are highly correlated with animal heart rate.
The median period length of cyclic tumor blood flow in this study was ~9 and 14 min in nude and C3H mice, respectively, which is in good agreement with period lengths detected in fluctuating blood flow or oxygenation in tumors of other investigations 
, including those in humans 
. Given our findings that these fluctuations correlate with changes in animal heart rate, it is interesting to note that this period is similar between tumors and/or strains as reported herein by us and in the observations of others 
because it supports the contribution of a more general-acting effector, such as heart rate, to fluctuation in tumor oxygenation or blood flow. Importantly, spontaneous fluctuations in murine heart rate with a similar period have been observed 
Strain-dependent differences in baseline hemodynamic function were evident in both cyclic fluctuations and longer-term trends. At the level of blood flow cycling, tumors of C3H had much more regular periods, evidenced by significantly higher autocorrelation functions than those found in the nudes. The C3H also were more likely to exhibit periodic fluctuations in tumor rBF and these fluctuations were completely absent from some of the nudes. In terms of longer term (over 1 hour) trends in tumor rBF, the C3H, but not the nudes, exhibited increasing blood flow over an hour of monitoring. Accompanying increases in heart rate were also detected in these animals. We suggest that these increases are related to the known vasodilative effect of isoflurane. Whereas we showed tumor rBF (C3H mice) to increase by ~25% in this study, Schumacher et al. reported isoflurane to induce a 26% increase in the diameter of microvessels of rat skeletal muscle 
. Neither tumor nor core animal temperature increased over this same monitoring period; moreover they did not differ between the strains, which shows that in the context of our investigations strain-dependent differences in tumor temperature did not exist or contribute to the differential in long term blood flow trends between the models.
Due to the anesthesia-induced increase in blood flow in the C3H animals, phosphorescence lifetime imaging was used to compare the oxygenation of tumors in C3H versus nude animals. Average (SD) tumor pO2 after 60 minutes of monitoring, i.e. the time point at which PDT would begin, was 20 (12) Torr in C3H and 33 (10) Torr in the nudes, showing substantial overlap between the strains. Thus, the stronger vascular effects of PDT in C3H mice could not be attributed to an improved state of oxygenation relative to the tumors in nude mice. A difference in Photofrin uptake between RIF tumors of nude and C3H animals was also ruled out as a cause of strain-dependent differences in PDT response.
In summary, comparative studies of the same tumor line grown in C3H and nude mice has found blood flow in the C3H animals to be more responsive to vascular stress, whether it be locally (PDT) or systemically (L-NNA) induced. Specifically, PDT uncouples cyclic fluctuations in tumor blood flow from animal heart rate, leading to decreases in tumor blood flow that are significantly greater in the C3H versus nude animals. These results may in part be explained by the smaller blood vessels of tumors in C3H mice, which could contribute to the general tighter control of blood flow in these tumors. All told, these results provide evidence that the underlying structure and hemodynamics of tumor blood vessels may inform upon the nature of their response to a subsequent vascular challenge. Differences in baseline tumor hemodynamics between animal strains should be considered when planning and interpreting studies of PDT, or other vaso-modulating applications.